Elka Popova, Research Manager, Frost & Sullivan
Q: Elka, you just came back from SpeechTEK. What did you find most surprising in your conversations with market participants?
A: The most interesting to see was how the buzzwords of yesteryear have become taboo today. For example, in 2004 the industry seemed really excited about packaged applications. At that time, everyone in the industry claimed packaged applications were going to drive mass adoption by reducing the total cost of ownership. To me it seemed obvious that some customization would always need to be done as no single application fits everybody's needs. But commoditizing frequently used modules that can be reapplied multiple times seemed like a great idea.
It is surprising, therefore, that this year people seemed to shudder at the mention of packaged applications. The reason? Some of them seem to have found out the hard way that speech applications can never become a true commodity. Others are really focused on leveraging their professional service expertise to grow revenues so they resent the concept of packaged applications.
In my view, application vendors should continue creating reusable modules and market simple applications, such as plug-and-play solutions, in order to simplify the sales process and reduce the total cost for their business customers. There is no reason to reject the concept of packaged applications only because we can't package all speech applications.
Q: What did you find most exciting and uplifting at the Spring Show?
A: The good news is that several market participants - ScanSoft and BeVocal to name a few - are actively hiring and that's always a good sign. Overall, the speech market seems to be growing, even though there hasn't been a growth spike yet. It does seem that the market will not follow a hockey stick curve, so there may never be an inflection point or "the year of speech."
I believe; however, that the speech industry is maturing in a number of different ways. On the demand side, customer awareness seems to be growing and businesses are becoming more sophisticated buyers. They are becoming increasingly proactive in seeking to deploy speech technologies and have more specific requirements regarding both the vendor and the speech application itself. You can tell this by how vendors are all striving to differentiate and find specific market niches where they can position themselves most competitively.
On the supply side, business models are gradually evolving to match vendors' core competencies and evolving market needs. Internal restructuring at vendors, such as ScanSoft, Nortel and IBM, shows that market participants are closely monitoring market trends and are adjusting their strategies accordingly. Even consolidation is a sign of a maturing industry, although this also has its down side. It takes care of the confusion over the availability of similar products from multiple vendors. Furthermore, it creates a more financially stable and viable industry.
Q: Where is speech being deployed today and what exciting developments does the future hold for speech vendors?
A: We are just now starting an in-depth study that will uncover specific opportunities for the deployment of speech technologies. It seems, however, that contact centers continue to represent the largest opportunity. Most of the exciting deployments of speech are in self-service solutions. Financial services including insurance companies, telcos and utility companies are among the verticals that continue to account for the majority of speech deployments. At the same time vendors are discovering new opportunities in other verticals such as doctors' offices, government offices and retail businesses.
Some of the more exciting future applications for the contact center market are likely to include video avatars targeted at the growing number of mobile subscribers accessing self-service applications over their mobile phones. Although this is probably a "nichy" application, as one industry pundit put it, it is likely to spice up customer service and help improve customer satisfaction. Multimodal applications will also gain traction once multimodal devices become widely available. Growing adoption of IP telephony and more sophisticated IP end devices as well as migration to 3G technologies in wireless networks will foster the adoption of multimodal applications.
Q: Are you seeing any growth potential for speech technologies in enhanced services offered by wireless or wireline carriers?
A: Enhanced services seem to account for an insignificant portion of speech deployments at present. Wireless carriers are currently entirely focused on data applications, and are trialing speech on a very limited scale. One of the challenges in the mobile service space is what Dr. K.W. "Bill" Scholz of Unisys calls "entitlement." Mobile subscribers consider themselves entitled to enhanced applications as part of their package and are reluctant to pay additional fees. Therefore, existing speech applications, such as voice-activated dialing and voice-navigated voicemail, are doomed to only limited penetration unless they become give-aways like voicemail, call waiting and caller ID.
Wireline carriers, on the other hand, are busy fighting off next-generation providers tapping into their customer bases. They are more focused on bundling local and long distance voice services at competitive monthly fees and integrating wireline and wireless services rather than looking to introduce new enhanced applications.
Neither wireless nor wireline carriers have found any compelling speech applications or compelling pricing and bundling structures for such applications yet. I believe; however, that, in the future, speech will become a widely accepted interface and a tool for improving customer convenience by providing customers with speech access to a variety of network-based features and capabilities through a speech portal. As speech-to-text technologies mature, they may be deployed to convert voice messages into short text messages (SMS) and gain some traction.
Q: Are there any specific factors that could positively impact market growth?
Numerous factors - company-specific, and industry-specific or general economic - can impact the speech market. I'll list just a few that strike me as particularly important.
A: The availability of tools or reusable modules/objects is likely to enable faster development and deployment of speech applications. Even though some vendors admit that they probably "oversold" this concept last year, I believe that tools and packaged applications will drive faster adoption by increasing the pool of application developers capable of developing speech applications, and by lowering the cost of speech solutions and shortening the development and deployment cycle.
Adoption of VoIP and IP telephony is likely to be another driver of speech deployment. As businesses and carriers migrate to a converged, next-generation infrastructure they can more cost-efficiently share resources across a distributed architecture. VoIP and VXML are enabling the merger of previously siloed applications, and are enabling businesses and carriers to realize significant cost savings. Furthermore, as businesses become increasingly virtual and modern professionals increasingly mobile, VoIP and speech technologies are likely to help businesses improve employee efficiency and productivity by giving employees access to converged voice and data applications anywhere anytime.
Finally, I think that hosted applications will help foster more rapid penetration of the small and medium business market. Hosted service providers such as BeVocal, Convergys, Qwest and Tellme offer their customers an opportunity to avoid the up-front investment in a rapidly evolving technology. By managing and maintaining the speech applications for their clients, they also relieve them of the need to develop in-house expertise. Hosted services also allow carriers to trial speech applications more rapidly and cost-efficiently and to introduce differentiating capabilities faster than competitors.
Q: What do you believe are some of the biggest threats lurking in the speech market?
A: There are various factors that could negatively impact market growth. Some represent a threat to multiple market participants, others - to individual vendors only. For example, extreme consolidation may restrict market penetration by slowing innovation and limiting price competition. Although this process may temporarily give an advantage to larger vendors, it will eventually hurt the growth prospects of the entire industry.
Similarly, slow and gradual speech adoption could prove detrimental to the viability of many vendors, but especially small ones focusing entirely on speech. Only tight cost control and a very sound business strategy can ensure the longevity of such market participants.
A clear and unbiased vision of the future of speech applications can help market participants develop sustainable business models. They should keep in mind that speech would be just another interface for a variety of information and communications applications. Web-based self service and GUI applications are not going to go away. Furthermore, customers will always prefer to talk to live agents no matter how sophisticated the self-service application is. Therefore, my advice to vendors is not to put all their eggs in the speech solutions basket.